| resumeang: Comments on the meta-analysis by M. A. Reinecke et al (see record 1997-38595-008) on the treatment of depression in young people. It is argued that the analysis may have overestimated the efficacy of cognitive-behavioral therapy (CBT). The article didn't cite 3 randomized studies of CBT that were published during the search period. All had negative results. It is appreciated that one of the studies was published in French and another in a British journal that is not widely available. There is evidence that negative findings are more likely to be published in such journals than in journals that are easier to access. Second, the strategy of including multiple measures from the same study as independent variables in the analysis, which leads to spurious confidence intervals, and hence to an overestimate of the statistical significance of the results. In addition, since the studies with the most outcome measures tended to have the largest effect sizes, the overall effect was probably inflated. (PsycINFO Database Record (c) 2000 APA, all rights reserved). |
